Integration of Existing Assets and Governance

NAISH does not start from zero. National use cases arrive with existing datasets, prior corpora, third-party models, governance frameworks, and DPI platforms that already work - and with inherited licences, consent terms, and access constraints that must be assessed before any of it touches the compute pipeline.

This section is the Hub's operating guidance for that reality: what already exists and how to adopt it, how to get a dataset from raw asset to pipeline-ready, and how to integrate with Nigeria's digital public infrastructure safely and repeatably.

The guidelines

How the three fit together

They follow the order in which a solution team actually meets them:

1. Search the catalogue before building. Reuse is the default on the Hub. A new build needs a reason, and an asset nobody catalogued is an asset the next team will rebuild.
2. Get the data ready. Whatever the catalogue provides, it enters the pipeline only through the licensing, pre-processing, bias, and versioning standards - and leaves it through reproducible, auditable training runs.
3. Integrate with DPI through the toolkit. Where a solution must touch national identity or other DPI, it does so through shared, tested components and playbooks, with the compliance evidence assembled before production access is requested from the platform owner.

Where this connects

• Feasibility - a use case cannot be judged technically feasible if its data assets are unlicensed, unprocessed, or unrepresentative, or if a DPI dependency has no viable access pathway.
• Data readiness gate - datasets that have not completed licensing review, pre-processing, and bias audit do not pass Data Readiness and are not admitted to the compute pipeline.
• Governance and risk - licensing terms, consent constraints, retention obligations, and DPI dependencies all feed the Governance & Risk framework.
• Responsible AI - bias audits and disaggregated, in-language evaluation are inputs to the Responsible AI pre-deployment gate.
• Prior work - the Projects Archive is where most of the catalogued assets come from.